Almost Timely News: 🗞️ The Dark Side of Generative AI (2025-02-23)

With great power comes great responsibility

Almost Timely News: 🗞️ The Dark Side of Generative AI (2025-02-23) :: View in Browser

The Big Plug

👉 Take my new course, Mastering Prompt Engineering for Marketers!

👉 Watch my latest talk, Generative AI for Tourism and Destination Marketing

Content Authenticity Statement

98% of this week's newsletter was generated by me, the human. There's a brief summary from Google Gemini in the opening section. Learn why this kind of disclosure is a good idea and might be required for anyone doing business in any capacity with the EU in the near future.

Watch This Newsletter On YouTube 📺

Click here for the video 📺 version of this newsletter on YouTube »

Click here for an MP3 audio 🎧 only version »

What's On My Mind: The Dark Side of Generative AI

How much harm is AI doing?

This is something of an existential question when it comes to AI. I’ve been in the space for 12 years now, from early days of R and machine learning back in 2013 until today. In that time, the field has exploded, from uplift modeling and attribution analysis to generative AI capabilities today that seem like science fiction.

But all this comes at a cost - and several costs. We’ve talked in the past about the sheer energy usage that AI demands, as much as 11% of all the electricity in the USA. But that's just one of the angles that we should examine if we're to evaluate AI fairly and understand its place and impact in our society.

Let's talk about the dark side of AI - what the many challenges are, and how we might overcome them.

Part 1: Intellectual Property

For years, there's been an ongoing battle between intellectual property holders - and that's all of us who create content (like this newsletter) - and AI model makers. The general substance of each side is that creators say AI infringes on their intellectual property rights because they were not compensated for the use of their data in the creation of AI models.

Moreover, AI models compete with those creators for commercial outcomes - a writer who is displaced by a generative AI model loses wages to that model. This was demonstrated in "AI and Jobs: Has the Inflection Point Arrived?", a paper from the National University of Singapore that showed earnings declined by 35% for translators once generative AI became fluent in their domain.

A second paper from Harvard Business School, "Who is AI Replacing? The Impact of GenAI on Online Freelancing Platforms" demonstrated a 21% decline in knowledge work jobs and a 17% additional decrease in graphic design jobs.

The model makers, the technology companies, argue that AI models are transformative works, that consuming our data and turning it into statistical libraries (which is what AI models are) is not making derivative works of our data. Transformative works are when a created work is so distinctly different from the original that no one could mistake it for the original. For example, painting rabbit ears on the Mona Lisa is clearly a derivative work. You can tell it's the Mona Lisa. If you took a print of the Mona Lisa, put it through a paper shredder and turned the shreddings into a sculpture of a rabbit, no one could reasonbly argue that the rabbit sculpture was recognizably the Mona Lisa.

On February 11, 2025, the US District Court ruled in Thomson Reuters v. Ross Intelligence that Ross Intelligence infringed on Thomson Reuters' copyrights to its headnotes when it turned them into training data for AI. What the court specifically ruled, summarized by Google Gemini, is:

The court rejected Ross's fair use defense. It found that the first fair use factor (purpose and character of use) weighed heavily against fair use because Ross's use was commercial and non-transformative. The court reasoned that Ross's use mirrored Thomson Reuters' purpose in using headnotes (legal research assistance) and did not add a new purpose or character. The court also found that the fourth factor (effect on the market) weighed against fair use because Ross's AI tool could potentially substitute for Westlaw and impact the market for training data using legal content. While the second factor (nature of copyrighted work) and third factor (amount and substantiality of portion used) favored Ross, the court determined that factors one and four were more critical and outweighed the others.

Now, the Ross model was not a general purpose AI model, like the ones that power ChatGPT and Google Gemini. It was purpose-built to compete directly with Thomson Reuters as a commercial substitute, but this is the first ruling that clearly says yes, training data - and thus the model - can infringe on copyright.

The argument that creators make - that AI competes in the marketplace with them - is proven in at least two studies (the ones above). A court ruling says AI intended to be a competitive replacement is an infringement of intellectual property rights.

So the first dark side of AI is pretty straightforward: models are infringing on intellectual property rights. Will this continue to hold up in court? It depends. In some parts of the world, this has already become established law. The EU AI Act requires any model maker of general purpose models to disclose the training data, and provides a remedy for intellectual property rights holders to either require withdrawing their content, or licensing it from them.

In other parts of the world, laws are different. In China, for example, intellectual property rights laws permit the use of source data to power AI models, so models like Qwen and DeepSeek are made in compliance with local laws. Whether or not they are ethical, they are legal in their jurisdictions. This poses a challenge to intellectual property rights globally because if one locale restricts the use of IP and another locale permits it, the permitting locale's models can advance faster and be more capable, generating a competitive advantage.

The remedy, if you're concerned about ethics and law, is to know where your models come from and what licensing the data has. For example, the French company Mistral's models are excellent. They don't outperform models like OpenAI o1 or DeepSeek R1 currently, but they are made in Europe and in accordance with the EU AI Act, so Mistral's models would fall under the category of ethically sourced. Non-technical people can use Mistral's Le Chat (their web interface) and more technical folks can download Mistral's models like Mistral Small 3 and run them on local hardware.

Part 2: Resource Costs

The second aspect of AI's dark side is resource consumption. AI consumes a lot of resources - electricity and fresh water in particular. OpenAI's new experimental model, o3, consumes more electricity for a high efficiency task than an entire single family home uses in 3 days. For its low efficiency tasks, it consumes more electricity than a single family home uses in a year and a half.

How much power are we talking about? If you go by the public estimates of how many GPUs NVIDIA has sold per year for the last 3 years, it looks like this - these are data center GPUs, not gaming cards:

• 2024: ~4.8M (guess)

• 2023: 3.76M

• 2022: 2.64M

• 2021: 1.58M

All of these will be A100 or H100 GPUs; the H100 came out in 2022 and got a refresh last year along with the H200 debut.

That’s 12.78 million GPUs. Export restrictions, especially to China (China and Taiwan make up about half of NVIDIA’s sales), started in 2022 and ramped up over the years. So call it half of those GPUs are likely in US data centers. Let’s make it 6 million for an even number, a little less than half. From here on out we'll be talking about the USA specifically; your mileage will vary based on your locale.

Every NVIDIA “GPU” is actually a 8 core blade. If you look at the product specs, they’ve had 8 cores since the A100. That means with 6 million GPUs, you’re talking 48 million cores. Each core uses 700 watts. That’s JUST the core of the GPU. An 8 core GPU consumes 5,600 watts.

So just on cores alone, you’re at 33.6 billion watts. (6 million GPUs * 8 cores each * 700 watts per core)

But we don’t use GPU cores, we use GPUs. They all need cooling and they all have heat waste. For example, the DGX H100 pod that has 8 H100 cores in it has a peak usage of 10,200 watts, an overhead power consumption of 4,600 watts above and beyond the cores themselves.

So 6 million GPUs * 4,600 watts (because we accounted for the core power already) is another 27.6 billion watts.

So the total draw is 61.2 billion watts, SOLELY for the GPUs. This doesn’t count running the actual data centers, the HVAC, etc. etc.

Now we have to convert watts to kWh, kilowatt hours. 61.2 billion watts is 61.2 million kilowatts. So 61.2 million kilowatts * the number of hours we care about is the kilowatt hours. If we’re talking energy use per year, there are 8,760 hours per year, so GPU usage is 536 billion kilowatt hours per year if you ran them full tilt, all the time.

We don’t know what percentage of GPUs are set aside for inference (aka regular use) and what percentage are dedicated to training. Those that are focused on training will be going full tilt, all the time. Those that are used for inference will fluctuate throughout the day. They’ll run hard during the workday and less hard when we’re all asleep. (Assuming that companies like Google have their GPUs distributed around the planet and not just in US data centers)

Now, the average household consumption of electricity in the USA for 142 million households at about 11,000 kWh per year is 1.54 trillion kWh for residential electrical use (which excludes commercial and industrial).

If you go for the absolute theoretical maximum - 100% GPU usage all the time, always and always, GPU usage could be 1/3 of US household electricity usage, or 11% of all electricity usage in the USA.

That's a HUGE amount of electricty at AI's theoretical maximum usage - and that's more carbon in the air.

And that doesn't count the fresh water needed to cool datacenters. In a 2024 paper from Carnegie Mellon University, "A Water Efficiency Dataset for African Data Centers" cited that a single GPT-4 task consumes up to 60 liters/15 gallons of water, based on electricity generation needs. Fresh water is consumed in two stages - onsite water for cooling, which uses as much as 10 liters of fresh water per 5,000 token (3,000 word) tasks, and offsite water as part of the generation of electricity itself.

The remedy here, if you're concerned about the environmental sustainability of generative AI, is to use local models that run on your consumer devices like laptops. Today's high-end laptops with great graphics cards can run AI models as big as 70 billion parameters, like Llama 3.3 or Qwen 2.5 72B. These models are very capable for many tasks and use only as much electricity as your laptop uses, and only as much fresh water as your electricity generation uses. At my house (and I completely acknolwedge my privilege here), I have solar panels on my roof. Most of my power use consumes no water at all, and my house is a net producer of green electicity about 6 months of the year.

Part 3: Labor Dislocation

The third dark side of AI is labor dislocation. We touched on this in part 1, but there's more to dig into. AI is displacing wages and as the technology improves, it will displace entire jobs, from the bottom up.

By that I mean the jobs AI will consume first in their entirety are entry-level jobs, jobs that require the lowest levels of skills and experience. For example, at the PR firm I used to work at, these were the tasks an account coordinator, the junior-most person on a PR team, would often do:

• Take notes during meetings and email them to the team and client

• First draft press releases

• Copy/paste news articles into a spreadsheet for team review (media monitoring)

• Copy/paste social media mentions into a spreadsheet

• Manually visit a client's social media profiles to count the number of followers the client has

• Make photocopies of slide decks and memos before meetings

• Email clients their reports

• Copy/paste journalist contacts from a central database to a spreadsheet (media list management)

• Fetch coffee for the managers

Other than the final task, there isn't a single task on this list that AI can't do faster, better, and cheaper. The role of account coordinator is largely replaceable by AI.

• Transcription software automatically takes notes, summarizes meetings, and distributes summaries

• Generative AI can generate press releases easily

• Basic automation and automation paired with language models (aka agents) can do ALL of the copy pasting

• Basic automation and automation paired with language models (aka agents) can browse social media profies as well as email reports

• Programmatic code can extract data from one database and put it in another database, and that code can be written entirely by AI

• Humans are still necessary to get coffee, but managers can also get their own damn coffee

Here's why this is a problem: junior-most roles are not only the easiest to replace in any organization with AI, they're also the most plentiful. For the profit-minded, this becomes massive savings because humans are expensive. Even the most junior role still requires a fair amount of overhead in many places (though more civilized countries with great social support reduce those costs considerably, places like Denmark, Norway, Finland, Sweden, Canada, Germany, France, etc.).

I spoke to a business owner not too long ago who said his key aim in 2025 was to shed 60% of his workforce that he lovingly called "dead weight" to boost his profitability and margins high enough to get acquired. That attitude is not uncommon, even if it's not pleasant to hear.

A recent paper in 2024 by the Tony Blair Institute for Global Change, "The Impact of AI on the Labour Market", the institute presents a number of different scenarios about AI's impact - and all of the scenarios (at least from my reading of the paper) are far too distant in the future. AI is accelerating at a pace unlike any other technology I've ever seen in my lifetime. In their fastest scenario, AI consumes up to 1.5 million jobs a year in the UK by 2040, or about 4% of the workforce per year.

Per year.

Humans are no stranger to economic dislocation. We've changed over the centuries from hunter-gatherers to agrarian societies, from farming to manufacturing, from making stuff to services-based economies, and today more and more of our economic activity is intangible in a knowledge economy. But past dislocations all happened over long periods of time. How you did work differed from how your kids and grandkids did work, and you never had to retrain in your lifetime.

Rapid change is what causes problems. Rapid economic shocks lead to things like civil unrest, even the fall of governments.

How rapid? Many papers have been written about social unrest from the COVID pandemic and other rapid dislocations. When unemployment jumps by more than 2% in a quarter and stays high in a society with substantial income inequality, social unrest and civic unrest are highly likely.

And the Blair Institute's looking at 4% per year, year after year. The math on that isn't hard to figure out.

The remedy here, to the extent there is one, is to heavily tax corporations and beneficiaries of AI (often nicknamed the robot tax), then use those proceeds to upskill citizens and provide a strong social safety net. At the individual and organizational level, it means investing heavily in your people, accepting some level of "inefficiency" and lower margins in exchange for keeping people productively employed rather than maximizing profit margins. It's in the best long-term interests of a society to do so to prevent social unrest and violent outcomes. Otherwise, as Louis XVI found out, angry poor people tend to bring out the guillotines.

Part 4: Enabling Bad Actors and Network States

The fourth and arguably darkest side of AI is the enablement of bad actors. I've said for years now that AI is an amplifier. Paraphrasing the Captain America movie, it takes the good and makes it better, and it takes the bad and makes it worse. Already, AI-powered scams have risen substantially, from simply fooling people into transferring hundreds of million dollars to election interference at scale.

But that's not the dark side we're talking about. In this case, we're talking about the system of AI being advanced by leading tech companies (mostly in the USA) and its consequences.

The first and most significant consequence is convincing everyone to use AI all the time, for everything, thus adding a basic tax on... well, pretty much everything. $20 a month per user for access to models and utilities like ChatGPT, plus building generative AI into pretty much everything. One look at the MarTech 14000 shows that nearly every company on the map is trying to put AI into their product somehow, if for no other reason than to be competitive in the marketplace.

On the surface, that doesn't seem that bad - except that all AI models have biases, without exception, and in some cases, model owners are making an attempt to intentionally bias model outputs. For example, in Grok's system prompts, there was a prohibition about avoiding answers that are "woke". If generative AI systems are the default source of information for people and we don't know what their biases are - intentional or otherwise - we will let them condition our thinking.

That leads us to the next challenge: where the money is going. We've already touched on income inequality as a source of social unrest, but AI magnifies it significantly more than other technologies. When a human does work, a human is paid wages. Those wages then are spent in the broader economy, from food to rent to enjoyable pastimes.

When a machine does work, the company owning the machine is paid instead, but the scale of labor is very different. A data center that can do the work of thousands of people typically employs about 50 people. Those 50 people are well compensated for their expertise, but the economics of that equation are going to cause problems. Most of the rest of the money goes to the development and sales of the technology.

Think about a company like Anthropic. It boasts about 88 million users a month, each doing tasks with Claude (which is an exceptional model family). Anthropic publishes Clio, its insights about how people use the service. Millions of people perform millions of tasks with Claude. How many people does Anthropic employ to deliver this service that helps substitute labor?

700.

Again, those 700 people earn a fantastic living for the services they provide, and the services they provide are clearly valuable. But 700 people cannot fuel a national economy.

Beyond that, there's a dystopian philosophy that underpins many in the technology community generally, from Marc Andreesen (of Netscape fame) to Vivek Ramaswamy to JD Vance to Peter Thiel (of Paypal fame) to Elon Musk: the Network State. The fundamental premise of the Network State is that current society (in general, and the USA in specific, though it's a global vision by its proponents) is broken and needs to be dismantled entirely, then rebuilt in alignment with venture capital ideas, led by tech leaders as an autocracy - like companies becoming sovereign nations.

This is the ultimate vision for some of these folks (who run AI companies, among other things): authoritarian regimes where they're the leaders making all the decisions, based on their beliefs that they're simply smarter than everyone else. The role AI plays? Substituting human labor at scale. The concept behind artificial general intelligence is pretty clear: machines that do things better than humans do. The endgame is also pretty clear: if you have the funding to do so, you can build a society of your choosing with most of the labor being done by machines. There's no pesky lower classes to overthrow you.

For most of us who aren't billionaires, this seems like a pretty raw deal, and it is. The challenge for many of us using generative AI to increase productivity and boost innovation is that our money is being used to fund these ideas which will ultimately be bad for us.

The remedy here should be no surprise: open model AI, like the models released by Mistral, DeepSeek, and so many other companies. These models are free to use; you download them, you run them, you absorb the operational cost, but in return, you pay nothing to people whose ultimate goals might literally involve your extinction.

There's a reason why the major tech companies and tech bros are spitting mad about DeepSeek in particular, spreading as much fear, uncertainty, and doubt as possible. Yes, there are VERY legitimate privacy concerns (or the complete lack of privacy) if you use DeepSeek's hosted models, which operate in the People's Republic of China. Their privacy policy for their services is pretty clear: they're using every bit of data you put into their services.

But their model is one that you can download onto your own hardware and run safely and privately. Companies like Amazon Web Services and Snowflake already let you run DeepSeek's models internally - and because Chinese companies (due to export restrictions) are compute-constrained, DeepSeek is very efficient and low-cost to operate.

Which deeply imperils the fortunes of the previously mentioned billionaire dictator wannabes. In a world where AI is a free utility, they can't economically sustain their nation states of the future. I spoke about this recently in Canada: every nation on the planet now, if it has the hardware to set up, can have state of the art AI for its citizens today simply by installing and running a local, safe, secure version of DeepSeek. Every nation on the planet could declare independence from mostly US-based technology providers - at the expense of the US-based technology billionaires.

Even more important, the innovations of companies like Mistral, Leilapa, DeepSeek, and so many others are often published in papers and code, enabling anyone who's skilled to use the technology for themselves. Unsloth recently released a toolkit that can convert any AI model into a DeepSeek-style reasoning model for less than $30 of compute time. A meal for two at McDonalds costs more now than tuning your own reasoning model - and that's good for everyone. It's especially good for economies and companies that don't have billions of dollars but have worthy goals.

Part 5: Wrapping Up

Reading through this is probably a little disheartening, because so much of the hype around AI is - correctly - about all the cool things it can do. I'm not saying to stop using AI, to lock it up. Like everything in life, there are tradeoffs. There are costs. There are dark sides, unintended consequences, and bad actors - and some of those bad actors make no effort to hide their bad actions, publishing them in manifestos like The Network State.

The purpose of this issue was to be honest about what the tradeoffs of AI are.

As I outlined in each section, for the most part, there are remedies. There are ways to offset most of the problems with AI, by using open models, by focusing on efficient use of the technology, by spending your money with the people and companies working to make the world as a whole better instead of making themselves richer to everyone else's detriment.

There are some fundamentally human problems, like bad actors, that are enabled by AI but are not solvable by AI. Those are only problems humans can solve for each other.

We live in more challenging, more uncertain times than ever before. The global world order is in flux. Nations around the planet are struggling to figure out what to do for... well, pretty much everything. But what we do have going for us is a set of tools - AI - that dramatically amplify our capabilities.

As Uncle Ben said to Peter Parker, with great power comes great responsibility. AI is a literal superpower. It's up to each of us, from the novice ChatGPT user to the system architects and model builders of AI, to use this power wisely and for the benefit of as many people as possible.

If we do that well, the future looks more like Star Trek and less like Mad Max.

How Was This Issue?

Rate this week's newsletter issue with a single click/tap. Your feedback over time helps me figure out what content to create for you.

• 😄 Good

• 😐 Neutral

• 😦 Bad

Here's The Unsubscribe

It took me a while to find a convenient way to link it up, but here's how to get to the unsubscribe.

If you don't see anything, here's the text link to copy and paste:

https://almosttimely.substack.com/action/disable_email

Share With a Friend or Colleague

If you enjoy this newsletter and want to share it with a friend/colleague, please do. Send this URL to your friend/colleague:

https://www.christopherspenn.com/newsletter

For enrolled subscribers on Substack, there are referral rewards if you refer 100, 200, or 300 other readers. Visit the Leaderboard here.

Advertisement: Bring Me In To Speak At Your Event

Elevate your next conference or corporate retreat with a customized keynote on the practical applications of AI. I deliver fresh insights tailored to your audience's industry and challenges, equipping your attendees with actionable resources and real-world knowledge to navigate the evolving AI landscape.

👉 If this sounds good to you, click/tap here to grab 15 minutes with the team to talk over your event's specific needs.

If you'd like to see more, here are:

• My speaker preview reel (YouTube)

• A full-length keynote you can enjoy

ICYMI: In Case You Missed It

This week, I did a very fun livestream on how to deconstruct AI use cases into something useful. Give it a watch!

• So What? How to Deconstruct Generative AI Use Cases

• Mind Readings: Sometimes Only Code Is The Answer

• Mind Readings: Treat Reasoning AI Models Like New Managers

• Mind Readings: Ask AI What You Can Improve

• Mind Readings: Sometimes Generative AI Is Not The Answer

• Almost Timely News: 🗞️ Solving Real Problems with Generative AI (2025-02-16)

• Tales of Economic Fear, TAM, Twitch and Smart Glasses!

• In-Ear Insights: Deconstructing Generative AI Use Cases

Skill Up With Classes

These are just a few of the classes I have available over at the Trust Insights website that you can take.

Premium

• 👉 New! Mastering Prompt Engineering for Marketers!

• 🦾 Generative AI for Marketers

• 👉 Google Analytics 4 for Marketers

• 👉 Google Search Console for Marketers (🚨 just updated with AI SEO stuff! 🚨)

Free

• New! Generative AI for Tourism and Destination Marketing

• Powering Up Your LinkedIn Profile (For Job Hunters) 2023 Edition

• Building the Data-Driven, AI-Powered Customer Journey for Retail and Ecommerce, 2024 Edition

• The Marketing Singularity: How Generative AI Means the End of Marketing As We Knew It

Advertisement: New AI Course!

Mastering Prompt Engineering for Marketers is a 2 hour tour through prompt engineering. The first couple of modules walk through not just what prompting is, but what's happening INSIDE the AI model as it processes a prompt. I made the explanation non-technical (because who really enjoys softmax layers and attention matrices besides me) but the walkthrough really digs into what's going on inside the box.

Knowing that helps us understand WHY prompts do or don't work. You'll see why in the course, when you watch how a prompt is processed.

Then we walk through 3 prompt frameworks, plus "delve" 😏 into advanced prompting techniques, along with a downloadable guide of what each technique is, why you should care, when you should use it, and how to use it.

After that, we get into knowledge blocks and priming representations, then how to build and manage a prompt library.

👉 Register here!

What's In The Box? Here's a 5 Minute Tour

Here's a 5 minute video tour of the course so you can see what's inside.

Get Back to Work

Folks who post jobs in the free Analytics for Marketers Slack community may have those jobs shared here, too. If you're looking for work, check out these recent open positions, and check out the Slack group for the comprehensive list.

• Chief Marketing Officer (Cmo) at HustleWing - Side hustles for professionals

• Chief Marketing Officer at Digna Recruitment

• Chief Marketing Officer at NextGen Coding Company

• Commerce Analytics Practice Lead at Verndale

• Digital Strategist at CyberSearch

• Google Ads Automation & Ai Specialist at Power Brands

• Head Of Marketing at Metric Bio

• Market Strategist at eSense Incorporated

• Marketing Gtm Growth Hacker at Lotus Interworks

• Marketing Lead at Plexus Resource Solutions

• Marketing Manager at Bio X Cell

• Sales Marketing Analyst at HRDelivered

• Senior Growth Marketer – B2b & B2c (Hr Tech & Employment Verification) at MyEmployment

• Sr. Performance Marketing Manager at Fractional Demand

Advertisement: Free Generative AI Cheat Sheets

Grab the Trust Insights cheat sheet bundle with the RACE Prompt Engineering framework, the PARE prompt refinement framework, and the TRIPS AI task identification framework AND worksheet, all in one convenient bundle, the generative AI power pack!

Download the bundle now for free!

How to Stay in Touch

Let's make sure we're connected in the places it suits you best. Here's where you can find different content:

• My blog - daily videos, blog posts, and podcast episodes

• My YouTube channel - daily videos, conference talks, and all things video

• My company, Trust Insights - marketing analytics help

• My podcast, Marketing over Coffee - weekly episodes of what's worth noting in marketing

• My second podcast, In-Ear Insights - the Trust Insights weekly podcast focused on data and analytics

• On Bluesky - random personal stuff and chaos

• On LinkedIn - daily videos and news

• On Instagram - personal photos and travels

• My free Slack discussion forum, Analytics for Marketers - open conversations about marketing and analytics

Listen to my theme song as a new single:

• Apple Music

• Spotify

• YouTube

Advertisement: Ukraine 🇺🇦 Humanitarian Fund

The war to free Ukraine continues. If you'd like to support humanitarian efforts in Ukraine, the Ukrainian government has set up a special portal, United24, to help make contributing easy. The effort to free Ukraine from Russia's illegal invasion needs your ongoing support.

👉 Donate today to the Ukraine Humanitarian Relief Fund »

Events I'll Be At

Here are the public events where I'm speaking and attending. Say hi if you're at an event also:

• Social Media Marketing World, San Diego, March 2025

• Content Jam, Chicago, April 2025

• TraceOne, Miami, April 205

• SMPS, Washington DC, May 2025

• SMPS, Los Angeles, Fall 2025

• SMPS, Columbus, August 2025

There are also private events that aren't open to the public.

If you're an event organizer, let me help your event shine. Visit my speaking page for more details.

Can't be at an event? Stop by my private Slack group instead, Analytics for Marketers.

Required Disclosures

Events with links have purchased sponsorships in this newsletter and as a result, I receive direct financial compensation for promoting them.

Advertisements in this newsletter have paid to be promoted, and as a result, I receive direct financial compensation for promoting them.

My company, Trust Insights, maintains business partnerships with companies including, but not limited to, IBM, Cisco Systems, Amazon, Talkwalker, MarketingProfs, MarketMuse, Agorapulse, Hubspot, Informa, Demandbase, The Marketing AI Institute, and others. While links shared from partners are not explicit endorsements, nor do they directly financially benefit Trust Insights, a commercial relationship exists for which Trust Insights may receive indirect financial benefit, and thus I may receive indirect financial benefit from them as well.

Thank You

Thanks for subscribing and reading this far. I appreciate it. As always, thank you for your support, your attention, and your kindness.

See you next week,

Christopher S. Penn